Evoluation of long ACL/EMNLP paper titles. Each column shows a cluster of years with similar content. Size indicates frequency; order characteristicness.

1979-1984	1985-1988	1989-1990	1991-1997	1998-1999	2001-2001	2002-2006	2007-2011	2012-2016	2017-2018	2019-2022
design	tree adjoining grammars	processor	plan	cross - language information	constraint - based	statistical language	semantic relatedness	short - term memory	neural semantic	bert
databases	sharing	lfg	connectionist	corpus - based	topological	wide - coverage	search queries	long short - term	attention model	pre - training
production	categorial	prepositions	lr	korean	text segmentation	lfg	phrase - based translation	distributional semantics	bandit	pre - trained
issues	tense	tree adjoining	unification - based	production	wordnet	maximum entropy	forest	distributional semantic	skip	transformers
computer	formalisms	unification	transducers	terminology	machine learning	field	web search	recursive neural	understand	few - shot
natural language interfaces	line	core	centering	using context	rule - based	model for statistical machine	system combination	community question answering	sequence models	pretrained
project	clauses	treatment	prepositional	routing	maximum entropy	words using	discriminative learning	abstract meaning representation	sequence learning	pre - trained language
access	massively	unification - based	semantic interpretation	ambiguities	treebank	using bilingual	forests	convolutional neural network	characters	multi - hop
speech acts	prosodic	representing	context - free grammars	processor	talk	large corpora	phoneme conversion	acl	lstms	self - supervised
linguistics	assignment	acquiring	ellipsis	automatic acquisition	fragments	machine learning	topic segmentation	recurrent neural networks	aware neural	contrastive learning